A novel Hybrid Machine learning approach (CNN-SVM) for COVID- ۱۹ diagnosis in CT images

This study aims to diagnose COVID-۱۹ using CT images and deep learning algorithms. First, we usewavelet transformation in combination with fuzzy logic to provide a new approach to removing the noiseof CT images. Then we segmented lung images by the proposed combined global and local thresholdmethod. In this way, lung regions from CT images can be segmented successfully. In the next step,features and classification will be extracted. AlexNet is used to extract features, while a Support VectorMachine (SVM) is used for classification. With ۹۹.۸% accuracy, three classes of data are classified:COVID-۱۹, Viral Pneumonia, and Normal. In comparison with previous methods, the proposed methodshows superior classification performance.

Keywords:

COVID-۱۹ , Convolutional neural networks , AlexNet , lung segmentation , Support vector machine (SVM)